too much for class?

what’s your success rate in using chatbots to do real work things you’ve tried them for (ie not “for fun” or “for testing [or “for a paper”]) where they’ve actually saved you time?

in the past month, they’ve saved me time this often:

every year around now i’m frustrated by the gross inefficiency (economic sense) of the CS faculty job market. it works-ish for the top 1% of departments and candidates and is awful for others. but without buyin from that 1% of departments it’s enormously difficult to change.

A study co-authored by Assistant Professor Jing Liu found racial disparities in disciplinary referrals. Based on these findings, Liu and his colleagues suggested that classroom-management training could impact racial discipline gaps.

Read more⬇️

wish i’d had this when i started!

two of my favorite researchers (and also favorite people) are looking for a postdoc!

Grateful for this thoughtful coverage of our upcoming #CSCW2024 work. Love how practically oriented it is.

Thankful to the participants without whom research isn't possible.

Congrats to the team Hal Daumé III Vera Liao Samir Passi Mark Riedl.

Microsoft Research Machine Learning at Georgia Tech

Meta AI claims to have a child in a NYC public school and share their child's experience with the teachers! The reply is in response to a question looking for personal feedback in a private Facebook group for parents. Also, Meta's algorithm ranks it as the top comment! AI at Meta

somehow only in teaching generative AI to freshmen without CS background did i realize that Reinforcement Learning from Human Feedback is actually RL from exactly NOT human feedback 🤷

i now get paid ads to learn about xrisk

what a world

ICYMI Mozilla assembled an impressive set of voices - inlc. folks at Meta, HuggingFace, Github, EleutherAI, Linux Foundation, etc - to discuss the definitions & difficulties of AI openness.

Worth checking out - I still regularly reference the tech memo:
blog.mozilla.org/en/mozilla/ai/…

Thrilled to announce that our workshop “Societal & User-Centered Privacy in AI (SUPA)” was accepted to #SOUPS2024 ! If you're interested in privacy in AI and how this impacts people & society, JOIN US

📅 : 23rd May 2024

More: tinyurl.com/Supa24

#SOUPS2024 #Privacy #AI

Snehesh Shrestha, a Ph.D. candidate at the Univ. of Maryland's department of computer science, is building bridges between AI and every other academic program on campus. wtop.com/maryland/2024/…

📣 NEW PAPER 📣 
Excited to share my first PhD paper accepted to NAACL 2024 Findings!
Can smaller #LLMs use fine-grained feedback to refine their translations? 🤔 🔎

arxiv.org/abs/2404.07851

🗣️🌏Submit your work on spoken language translation to IWSLT by April 29!
In addition to shared task papers, you can submit regular scientific papers. And we have a special highlights track to reflect on emerging trends, or share work published elsewhere with the IWSLT community

whole thread but what really resonated was how weird it is that in DV circles open weights are seen as bad … cuz that’s so divorced from what i see as general academic agreement. no open weight models also dooms a lot of academic research.

A really fantastic job opportunity at OMB just opened up today — this person would be central to the White House's oversight of federal agencies' use of AI: usajobs.gov/job/786286500

We're excited to share the first CFP for CSCW 2025 with a new paper submission deadline of July 2, 2024. Read for more details and share with your networks: cscw.acm.org/2025/ #cscw2025

(1/N) New paper! Dataset Reset Policy Optimization for RLHF (arxiv.org/pdf/2404.08495…)

RLHF is a popular paradigm for fine-tuning generative models. But the question is, can we design algorithms that take advantage of additional properties of the RLHF framework?

📢 I'm recruiting a postdoc at UMD
🗓️ Starting in 2024
🖥️🙋 Topics related to human-AI interaction
🔜 Apply by May 31, but earlier if possible!
🤑 $75-$80k+benefits, unencumbered (ie not 'big project') funding

Link to full post and form for applications in thread.

>